Interpretable
Intelligence

Prologue

0.The Priests of AGI

I · Foundations

1.What Is Interpretability?

2.The Underdetermination of Intelligence

3.What Can an Explanation Tell Us?

II · Methods

4.Input Attribution

5.Concepts and Representations

6.Training Data

III · Learning aControl

7.Learning with Explanations

8.Interpretability Efficiency and Scaling

9.Interpretable Agents

IV · Systems

10.Generative World Models

11.Generative Language Models

12.Interpretable AlphaFold

13.Interpretable Chess and GRPO

V · Open Problems

14.Open Problems

Learning and Control

Chapter 8

Interpretability Efficiency and Scaling

Measuring interpretability by cost-to-target and studying how its value scales.

Coming soon.

← PreviousLearning with Explanations

Next →Interpretable Agents

Interpretable Intelligence synthesizes work from across the field, including research from Guide Labs. Relevant authors and results are cited throughout.

About this collection·Contents